7 Understanding the Oracle Real Application Clusters Installed Configuration

This chapter describes the configuration of an Oracle Real Application Clusters (Oracle RAC) installation.

This chapter contains the following topics:

7.1 Understanding the Oracle RAC Configuration

Oracle Net Configuration Assistant (NETCA) and Oracle Database Configuration Assistant (DBCA) configure your environment to meet the requirements for database creation and to enable discovery of Oracle RAC databases by Oracle Enterprise Manager.


Configuration files are created on each node of your cluster database.

Avoid changing host names after you complete the Oracle RAC installation, including adding or deleting domain qualifications. Node names are created from the host names during an Oracle Clusterware installation and are used extensively with database processes. Nodes with changed host names must be deleted from the cluster and added back with the new host names.

7.2 Understanding Time Zone Settings on Cluster Nodes

Oracle RAC requires that all cluster nodes have the same time zone setting. During an Oracle Clusterware installation, the installation process determines the time zone setting of the owner of the Oracle Grid Infrastructure installation on the node where Oracle Universal Installer (OUI) runs, and uses that time zone value on all of the nodes as the default time zone setting for all processes that Oracle Clusterware manages. This default setting is used for databases, Oracle Automatic Storage Management (Oracle ASM), and any other managed processes.

If you use SQL*Plus to start an instance, then you must ensure that the time zone value used by Oracle RAC is the same as the time zone value configured in Oracle Clusterware for the database. You can change the time zone that Oracle Clusterware uses for a database by running the following command, where time zone is the time zone to which you want to change:

srvctl setenv database -T "TZ=time zone"

7.3 Working with Multiple Oracle Home Directories on Windows

Each Oracle product should be installed in its own Oracle home. In other words, each product should be installed into a different directory structure from other Oracle products. The value for %ORACLE_BASE% is stored in the registry (for example, in HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\HOME0). The values for ORACLE_HOME and ORACLE_SID are also stored in the registry. Symbolic links for these directories, like those used on UNIX platforms, are not supported on Windows platforms.

See Also:

Oracle Database Platform Guide for Microsoft Windows for more information about the registry keys

Oracle Universal Installer (OUI) records the values for environment variables such as ORACLE_BASE, ORACLE_HOME, and ORACLE_SID in the registry and also updates the value for the PATH environment variable for the user performing the installation. In Linux and UNIX systems, you must manually set these environment variables in the user session or user profile.

To change the current setting for %ORACLE_HOME% (the default Oracle home), perform the following steps:

  1. Run OUI, release 10.n or later.

  2. Click the Installed Products button.

  3. Click the Environment tab at the top of the window.

  4. Move the Oracle home directory that you want as your default to the top of the list.

  5. Apply the changes, and exit the installer.

This procedure changes the value of the default ORACLE_HOME variable in the registry to the value you selected. It also ensures that the %ORACLE_HOME%\bin directories for each product are listed in the correct order in your PATH environment variable.

7.4 Understanding the Database Components Created by DBCA

The database components that DBCA creates include:

7.4.1 Tablespaces and Data Files

In both single-instance and cluster database environments, an Oracle database is divided into smaller logical areas of space known as tablespaces. Each tablespace corresponds to one or more data files on the shared storage. Table 7-1 shows the tablespace names used by an Oracle RAC database and the types of data they contain.

Table 7-1 Tablespace Names Used with Oracle Real Application Clusters Databases

Tablespace Name Contents


A mandatory tablespace that consists of the data dictionary, including definitions of tables, views, and stored procedures needed by the database. Oracle Database automatically maintains information in this tablespace.


A mandatory, auxiliary system tablespace that is used by many Oracle Database features and products. This tablespace contains content that was previously stored in the DRSYS, CWMLITE, XDB, ODM, OEM_REPOSITORY, and SYSTEM tablespaces.


An user-created tablespace that consists of application data. As you create and enter data into tables, Oracle Database fills this space with your data.


A mandatory tablespace that contains temporary tables and indexes created during SQL statement processing. You may have to expand this tablespace if you run SQL statements that involve significant sorting, such as ANALYZE COMPUTE STATISTICS on a very large table, or the constructs GROUP BY, ORDER BY, or DISTINCT.


System-managed tablespaces that contain undo data for each instance. Each Oracle RAC instance uses a different value for n in the tablespace name. These tablespaces are used for automatic undo management.


A system tablespace that contains rollback segments. If you do not use automatic undo management, then you must configure the RBS tablespace. The RBS tablespace should only be used when needed for compatibility with earlier versions of Oracle Database.

See Also:

Oracle Database Administrator's Guide for more information about the SYSTEM, SYSAUX, and other tablespaces

You cannot alter these tablespace names if you select a preconfigured database templates, or database configuration type when installing Oracle RAC. However, you can change the names of the tablespaces if you select the Advanced database configuration type during the Oracle RAC installation. See "Selecting an Oracle Database Configuration Type" for more information about database configuration types.

As mentioned, each tablespace has one or more data files. The data file names created by the preconfigured database configuration types vary by storage type such as Oracle Automatic Storage Management (Oracle ASM), or a file system.

Windows platforms use symbolic link names for the data files and other database files as shown in Table 7-2. You can specify different symbolic link names when you select the Advanced database configuration type during installation.

Table 7-2 Symbolic Link Names for Database Files on Windows Platforms

Windows Symbolic Link Name Tablespace or Other Database File Name






















First control file


Second control file


Where instance_number is the number of the instance and log_number is the redo log number (1 or 2).

Online redo log files

Each database instance must have at least two online redo log files. If the database name is db, then the symbolic link names for online redo log files for the first database instance are:



The link names for the online redo log files for the second instance are:



7.4.2 Control Files

The database is configured with two control files that must be stored on shared storage. Every database must has one unique control file; any additional control files configured for the database are identical copies of the original control file.

If a control file becomes unusable, then the database instance fails when it attempts to access the damaged control file. By multiplexing (creating multiple copies of) a control file on different disks, the database can achieve redundancy and thereby avoid a single point of failure.

See Also:

Oracle Database Concepts for more information about the control file

7.4.3 Online Redo Log Files

Each database instance must have at least two online redo log files. The online redo log files for a database instance are called the redo thread. Each Oracle RAC database instance has its own redo thread to avoid contention for a single set of online redo log files. In case of instance failure, the online redo log files must be accessible by the surviving instances. Therefore, the online redo log files for an Oracle RAC database must be placed on shared storage or Oracle ASM. If you use a file system for storage, then the file system must be a shared or cluster file system.

The generated file names created by the preconfigured database configuration types for the online redo log files can vary depending on the type of storage used. Unless you are using a cluster file system or Oracle ASM, during installation or database creation, you must specify a raw device name for the storage location of each online redo log file.

See Also:

7.5 Managing Undo Tablespaces in Oracle RAC

Oracle Database stores rollback or undo information in undo tablespaces. To manage undo tablespaces, Oracle recommends that you use automatic undo management, which is an automated management mode for the undo tablespace that makes the undo tablespace easier to administer.

When Oracle ASM and Oracle Managed Files (OMF) are used along with automatic undo management, an instance that is started for the first time, and thus does not have an undo tablespace, has its undo tablespace created for it by another instance automatically. The same is also true for online redo logs.

See Also:

7.6 Using Initialization Parameter Files

Oracle recommends that you use a server parameter file (SPFILE) for storing Oracle Database initialization parameters. Oracle recommends that you store all SPFILEs on Oracle ASM, including the Oracle ASM SPFILE. SPFILEs must be located on shared storage so that all instances in a cluster database can access this parameter file.

See Also:

Chapter 5, " Configuring the Server Parameter File in Oracle Real Application Clusters Environments" for more information about the creation and use of parameter files

7.7 Oracle Database Vault Default Audit Policy and Initialization Parameters

Oracle Database Vault installs a baseline database auditing policy. This policy covers the access control configuration information stored in Oracle Database Vault tables, information stored in the Oracle data dictionary (rollback segments, tablespaces, and so on), the use of system privileges, and Oracle Label Security configuration. When you install Oracle Database Vault, the security specific database initialization parameters are initialized with default values.

See Also:

Oracle Database Vault Administrator's Guide for more information about how installing Oracle Database Vault affects the Oracle RAC database configuration

7.8 Understanding Oracle Net Services Configuration Files

Networking elements for the Oracle Database server and clients are preconfigured for most environments. The Easy Connect naming method is enabled by default and does not require a repository. If you use a naming method other than Easy Connect, then additional configuration of Oracle Net Services may be required.

The following sections the Oracle Net Services configuration for an Oracle RAC database:

7.8.1 Database Services for an Oracle RAC Database

Each database is represented by one or more services. A service is identified by a service name, for example, sales.example.com. A client uses a service name to identify the database it must access. The information about the database service and its location in the network is transparent to the client.

The information needed to use a service name to create a database connection is stored in a repository, which is represented by one or more naming methods. A naming method is a resolution method used by a client application to resolve a connect identifier (such as the service name) to a connect descriptor when attempting to connect to a database service. Oracle Net Services offers several types of naming methods that support localized configuration on each client, or centralized configuration that can be accessed by all clients in the network.

An Oracle Database 11g release 2 (11.2) database service automatically registers with the listeners specified in the LOCAL_LISTENER and REMOTE_LISTENER parameters. During registration, PMON sends information such as the service name, instance names, and workload information to the listeners. This feature is called service registration

Services coordinate their sessions by registering their workload, or the amount of work they are currently handling, with the local listener and the SCAN listeners. Clients are redirected by the SCAN listener to a local listener on the least-loaded node that is running the instance for a particular service. This feature is called load balancing. The local listener either directs the client to a dispatcher process (if the database was configured for shared server), or directs the client to a dedicated server process.

When a listener starts after the Oracle instance starts, and the listener is available for service registration, registration does not occur until the next time the Oracle Database process monitor (PMON) starts its discovery routine. By default, the PMON discovery routine is started every 60 seconds. To override the 60-second delay, use the SQL statement ALTER SYSTEM REGISTER. This statement forces PMON to register the service immediately.


Oracle recommends that you create a script to run this statement immediately after starting the listener. If you run this statement when the instance is registered are services are currently registered, or while the listener is down, then the statement has no effect.

See Also:

Oracle Database Net Services Administrator's Guide for more information about service registration

7.8.2 Configuring the Listener Configuration File (listener.ora)

You can configure Oracle Net Services and listeners as described in the following sections: Grid Naming Service and Listeners

If you enable GNS, then you do not have to manually configure the listener.

In GNS configurations, there is one GNS daemon for the cluster. Service requests to the cluster domain that GNS manages are routed to the GNS VIP address, which routes these requests to the GNS daemon.

The GNS daemon listens for registrations. When a SCAN VIP starts on a node, it registers its addresses with GNS. When GNS receives a request from a DNS for the SCAN, it returns the registered addresses to the DNS. Local Listeners

Starting with Oracle Database 11g release 2 (11.2), the local listener, or default listener, is located in the Grid home when you have Oracle Grid Infrastructure installed. The listener.ora file is located in the Grid_home\network\admin directory. You can use a non-default location for the listener.ora file by setting the TNS_ADMIN environment variable or registry value to point to the directory that contains the Oracle Net Services configuration files.

During Oracle Database creation, the LOCAL_LISTENER parameter is automatically configured to point to the local listener for the database. The Database Agent process (oraagent.exe, previously known as racgimon) sets the LOCAL_LISTENER parameter to a connect descriptor that does not require an Oracle Net service name.

You can set a value manually for LOCAL_LISTENER. If you set LOCAL_LISTENER, then the Database Agent process does not automatically update this value. Oracle recommends that you leave the parameter unset so that the Database Agent process can maintain it automatically. If you do not set LOCAL_LISTENER, then the Database Agent process automatically updates the database associated with the local listener in the Grid home, even when the ports or IP address of that listener are changed.

See Also: How Oracle Database Uses the Listener File (listener.ora)

Local listeners are configured to respond to database connection requests, and to nondatabase connection requests, such as external procedures or Oracle XML Database (XDB) requests. The listener.ora file is the configuration file for a listener. It can include the protocol addresses it is accepting connection requests on, a list of the database and nondatabase services it is listening for, and control parameters used by the listener. You can modify the configuration of the listeners used by Oracle Clusterware and Oracle RAC with Server Control Utility (SRVCTL) commands, or by using NETCA. Manual editing of the listener.ora file is not required.

Before you install Oracle RAC, during the Oracle Grid Infrastructure installation, NETCA creates and starts a default listener in the Grid home called LISTENER. The listener is configured with default protocol listening addresses. The listener is configured to respond to connection requests that are directed to one protocol address specified during installation.

During the Oracle RAC installation, the Oracle RAC database uses the listener in the Grid home, and configures service information about the Oracle RAC database. The database services are registered automatically and do not require configuration in the listener.ora file. Dynamic service registration eliminates the need for static configuration of database services. However, static service configuration is required if you plan to use Oracle Enterprise Manager.

Each listener is configured with one or more protocol addresses that specify its listening endpoints. The protocol address defines the protocol the listener listens on, and any other protocol specific information, such as the address, host, and port. Starting with Oracle Database 11g release 2, the listener.ora file now contains only an IPC key and the following information:


In the previous example, the protocol ADDRESS refers implicitly to the HOST endpoint of the local node. The listener.ora file is the same on every node for an Oracle RAC database. Listening endpoints, such as the port numbers, are dynamically registered with the listener.

Example 7-1 Example listener.ora File for an Oracle RAC Node

The following is an example listener.ora file as it would appear after installation, with an entry for a node named node1 and a SCAN listener.

SCAN1))))               # line added by Agent
          # line added by Agent
# listener.ora.mycluster Network Configuration File:
# Generated by Oracle configuration tools.
ENABLE_GLOBAL_DYNAMIC_ENDPOINT_LISTENER_SCAN1=ON              # line added by Agent Configuring Multiple Listeners for Your Database Using DBCA

It is possible to configure multiple listeners, each with unique name, in one listener.ora file. Multiple listener configurations are possible because each of the top-level configuration parameters has a suffix of the listener name or is the listener name itself.


Oracle recommends running only one listener for each node in most customer environments.

You can use DBCA to configure which listeners your Oracle RAC database registers with. If DBCA detects more than one listener on a node, then it displays a list of the listeners. You can select one or all of these listeners with which to register your database.


To administer Oracle Database 11g release 2 (11.2) local and SCAN listeners using the lsnrctl command, set your ORACLE_HOME environment variable to the path for the Grid home. Do not attempt to use the lsnrctl commands from Oracle home locations for earlier releases, because they cannot be used with Oracle Database 11g release 2 (11.2).

7.8.3 Net Service Names (tnsnames.ora File)

A tnsnames.ora file is created on each node and acts as a repository of net service names. Each net service name is associated with a connect identifier. A connect identifier is an identifier that maps a user-defined name to a connect descriptor. A connect descriptor contains the following information:

  • The network route to the service, including the location of the listener through a protocol address

  • The SERVICE_NAME parameter, with the value set to the name of a database service


    The SERVICE_NAME parameter you use in the tnsnames.ora file is singular, because you can specify only one service name. There is also the service_names parameter in the database initialization file. The service_names database parameter defaults to the global database name, a name comprising the db_name and db_domain parameters in the initialization parameter file.

Example 7-2 Example tnsnames.ora File

The following is an example of a tnsnames.ora file. It contains a single net service name entry. The connect identifier for the net service name is sales. The database service associated with the sales connect identifier is myApp. The connect descriptor is the information in the line that starts with ADDRESS.

#  tnsnames.ora Network Configuration File:
#  Generated by Oracle configuration tools.

sales =
   (ADDRESS = (PROTOCOL = TCP)(HOST = myscan.mycluster.example.com)(PORT = 1522))
        (SERVICE_NAME = myApp)

The tnsnames.ora file is located in both the Grid_home\network\admin and Oracle_home\network\admin directories. By default, the tnsnames.ora file is read from the Grid home when Oracle Grid Infrastructure is installed.

Example 7-3 Example tnsnames.ora Entry for an Oracle RAC Database

The following is an example of a net service name stored in the tnsnames.ora file for an Oracle RAC 11g release 2 (11.2) database after installation, where the database name is RACDB, the cluster name is mycluster, and GNS is configured for the cluster:

     (ADDRESS = (PROTOCOL = TCP)(HOST=myscan.mycluster.example.com)(PORT=1521))

With Oracle Clusterware 11g release 2 and later, the listener association no longer requires tnsnames.ora file entries. The listener associations are configured as follows:

  • DBCA no longer sets the LOCAL_LISTENER parameter. The Oracle Clusterware agent that starts the database sets the LOCAL_LISTENER parameter dynamically, and it sets it to the actual value, not an alias. So listener_alias entries are no longer needed in the tnsnames.ora file.

  • The REMOTE_LISTENER parameters is configured by DBCA to reference the SCAN and SCAN port, without any need for a tnsnames.ora entry. Oracle Clusterware uses the Easy Connect naming method with scanname:scanport, so no listener associations for the REMOTE_LISTENER parameter are needed in the tnsnames.ora file.

See Also:

Oracle Database Net Services Administrator's Guide for more information about the tnsnames.ora file

DBCA creates net service names for connections as described in the following sections: Net Service Names for Database Connections

Clients that connect to any instance of the database should use SCAN in the connect descriptor. You could also use a net service name to connect to the database. The net service name created by DBCA enables Oracle Enterprise Manager to discover an Oracle RAC database.

Three SCAN addresses are configured for the cluster, and allocated to servers. When a client issues a connection request using SCAN, the three SCAN addresses are returned to the client. If the first address fails, then the connection request to the SCAN name fails over to the next address. Using multiple addresses allows a client to connect to an instance of the database even if the initial instance has failed.

Example 7-4 Sample Net Service Name Entry for a Database Connection

This example shows a connect descriptor that is used in a tnsnames.ora file. The connect identifier in this case is the same as the database name, mycluster.example.com. Instead of specifying an individual server, virtual Internet Protocol (VIP) address, or cluster node name, the connect descriptor uses SCAN, which is myscan.mycluster.example.com.

Connections that use the net service name mycluster.example.com are connected to any of the database instances of the mycluster database that runs the myApp database service.

mycluster.example.com =
    (ADDRESS = (PROTOCOL = TCP)(HOST = host=myscan.mycluster.example.com)
      (PORT = 1522))
      (SERVICE_NAME = myApp)

The net service name does not need to know the physical address of the server on which the database, database instance, or listener runs. SCAN is resolved by DNS, which returns three IP addresses to the client. The client then tries each address in succession until a connection is made.

Example 7-5 Connecting to Oracle RAC Using the Easy Connect Naming Method

Clients can use SCAN and the Easy Connect naming method to connect to an Oracle RAC database without configuring the tnsnames.ora file. If the Oracle RAC database runs on a cluster for which SCAN is myscan.mycluster.example.com, then you could use a connection request for the database service RACDB.example.com using a connect descriptor similar to the following:

username/@myscan.mycluster.example.com:1521/RACDB.example.com Net Service Name for Instance Connections

Clients that connect to a particular instance of the database use the net service name for the instance. The tnsnames.ora entry in Example 7-6 enables Oracle Enterprise Manager to discover the database instances in the cluster. These entries are also used to start and stop instances.

Example 7-6 Sample Net Service Name Entry for an Instance Connection

This example shows a net service name that is used to connect to a specific database instance. The connect identifier is the same as the instance name, mycluster1.example.com. The connect descriptor uses SCAN to locate the instance. Connections that use the net service name mycluster1.example.com are connected to the mycluster1 database instance of the mycluster database, without needing to know on which node the instance is running.


7.8.4 Net Services Profile (sqlnet.ora File)

OUI starts NETCA after the database is created. NETCA creates the Oracle Net Services profile, or the sqlnet.ora file. In an Oracle Grid Infrastructure installation, the sqlnet.ora file is located in the following directory by default:


In the %ORACLE_HOME%\network\admin directory, in addition to a default sqlnet.ora file, you also can find a sample sqlnet.ora file in the subdirectory sample.

During installation, NETCA creates the following entries in the sqlnet.ora file, where %ORACLE_BASE% is the path to the Oracle base directory for the Oracle RAC installation:


The AUTHENTICATION_SERVICES parameter specifies the method by which users are authenticated for database access. The value NTS indicates that Microsoft Windows native operating system authentication should be used to authorize access to the database. The parameter NAMES.DIRECTORY_PATH specifies the priority order of the naming methods to use to resolve connect identifiers to connect descriptors. The ADR_BASE parameter specifies the base directory into which tracing and logging incidents are stored when Automatic Diagnostic Repository (ADR) is enabled for the database.

The Easy Connect naming method eliminates the need to lookup service names in the tnsnames.ora files for TCP/IP environments. With Easy Connect, clients use a connect string for a simple TCP/IP address, consisting of a host name and optional port and service name. If you use this method, then no naming or directory system is required. See Example 7-5, "Connecting to Oracle RAC Using the Easy Connect Naming Method" for an example.

See Also:

7.9 Using SCAN with Oracle RAC Databases

Oracle RAC databases provide the important benefits of connection load balancing and failover. Standalone Oracle databases perform load balancing by distributing connections among the shared server dispatcher processes.

Oracle RAC databases increase load balancing by using the single client access name (SCAN) for the cluster to balance connections among all instances in a cluster database. SCAN also adds location independence for the databases, so that client configuration does not have to depend on which nodes are running a particular database or database instance.

Review the following sections to understand Oracle RAC load-balancing features:

7.9.1 Understanding SCAN

SCAN is a fully qualified name (host name.domain name) that is configured to resolve to all the addresses allocated for the SCAN listeners. The default value for SCAN is cluster_name.GNS_sub_domain, or, cluster_name-scan.domain_name if GNS is not used. For example, in a cluster that does not use GNS, if your cluster name is salesRAC, and your domain is example.com, then the default SCAN address is salesRAC-scan.example.com:1521.

SCAN is configured in DNS to resolve to three IP addresses, and DNS should return the addresses using a round-robin algorithm. This means that when SCAN is resolved by DNS, the IP addresses are returned to the client in a different order each time.


Configuring SCAN in DNS using a round-robin algorithm does not ensure failover of connections. Instead, failover of connections is handled by the Oracle Client. Oracle recommends that you use Oracle Client 11g release 2 for connections that use SCAN.

SCAN provides location independence for the databases, so that the client configuration does not have to include the address of each node that hosts a particular database or database instance. For example, if you configure policy-managed server pools for a cluster, then SCAN enables connections to databases in these server pools regardless of which nodes are allocated to the server pool.

During the installation of Oracle Grid Infrastructure, several Oracle Clusterware resources are created for SCAN:

  • A SCAN VIP is created for each IP address that SCAN resolves to

  • A SCAN listener is created for each SCAN VIP

  • A dependency on the SCAN VIP is configured for the SCAN listener

SCAN VIPs function like node VIPs, except that SCAN VIPs can run on any node in the cluster. Also, if you have three SCAN VIPs but only two nodes in your cluster, you can have two SCAN VIPs running on the same server. Clients (users or applications) that connect using SCAN instead of the node VIPs do not have to update the list of VIP addresses in their local tnsnames.ora file when nodes are added to or removed from the cluster.

7.9.2 Understanding SCAN Listeners and Listener Registration

During Oracle Grid Infrastructure installation, SCAN listeners are created for as many IP addresses as there are addresses assigned to resolve to the SCAN. Oracle recommends that the SCAN resolves to three addresses, to provide high availability and scalability. If the SCAN resolves to three addresses, then there are three SCAN listeners created.

The addresses for the SCAN listeners resolve with either the domain name service (DNS), or within the cluster for the Grid Naming Service (GNS), using a round-robin method. SCAN listeners can run on any node in the cluster.

The database parameter LOCAL_LISTENER specifies the listening endpoint of the local database listener, and the database parameter REMOTE_LISTENER parameter identifies the SCAN listeners. The database registers with the local and SCAN listeners by using the connect description information contained in these parameters. Oracle Database 11g release 2 and later instances only register with SCAN listeners as remote listeners. Upgraded databases register with SCAN listeners as remote listeners, and also continue to register with all node listeners.

Starting with Oracle Database 11g release 2, the REMOTE_LISTENER parameter is always set to the SCAN address. Do not set it to an Oracle Net alias that has a single address that uses SCAN for the host name (HOST=scan). For example, if SCAN for the cluster is myscan, and the GNS subdomain for the cluster is mycluster.example.com, then the REMOTE_LISTENER parameter has the following value:


The Database Agent process (previously known as racgimon) configures the LOCAL_LISTENER initialization parameter automatically for Oracle RAC databases by specifying a connect descriptor for the node VIP address of the local node. You do not have to alter this setting. The REMOTE_LISTENER parameter should be set to the SCAN and SCAN port. If you change the SCAN port, then you must change the value for the REMOTE_LISTENER parameter using the SQL statement ALTER SYSTEM SET REMOTE_LISTENER.

7.9.3 Understanding Failover, Load Balancing, and SCAN

Oracle RAC provides failover with the node VIP addresses by configuring multiple listeners on multiple nodes to manage client connection requests for the same database service. If a node fails, then the service connecting to the VIP is relocated transparently to a surviving node, enabling fast notification of the failure to the clients connecting through the VIP. If the application and client are configured with transparent application failover options, then the client is reconnected to the surviving node.

Clients configured to use VIP addresses for Oracle RAC releases prior to Oracle Database 11g release 2 can continue to use their existing connection addresses; using SCAN is not required. When an earlier release of Oracle Database is upgraded, it registers with the SCAN listeners, and clients can start using SCAN to connect to that database.

There are two types of load balancing that you can implement for an Oracle RAC database: client-side and server-side load balancing. Client-side load balancing balances the connection requests across the listeners. With server-side load balancing, the SCAN listener directs a connection request to the best instance currently providing the service by using the load balancing advisory.

Based on the environment, the following actions occur when you use SCAN to connect to an Oracle RAC database using a service name. The numbered actions correspond to the arrows shown in Figure 7-1.

  1. The PMON process of each instance registers the database services with the default listener on the local node and with each SCAN listener, which is specified by the REMOTE_LISTENER database parameter. The listeners are dynamically updated on the amount of work being handled by the instances and dispatchers.

  2. The client issues a database connection request using a connect descriptor of the form:



    If you use the Easy Connect naming method, then ensure the sqlnet.ora file on the client contains EZCONNECT in the list of naming methods specified by the NAMES.DIRECTORY_PATH parameter.
  3. The client uses DNS to resolve scan_name. After DNS returns the three addresses assigned to SCAN, the client sends a connect request to the first IP address. If the connect request fails, then the client attempts to connect using the next IP address.

  4. When the connect request is successful, the client connects to a SCAN listener for the cluster which hosts the sales database. The SCAN listener compares the workload of the instances sales1 and sales2 and the workload of the nodes on which they are running. Because node2 is less loaded than node1, the SCAN listener selects node2 and sends the address for the listener on that node back to the client.

  5. The client connects to the local listener on node2. The local listener starts a dedicated server process for the connection to the database.

  6. The client connects directly to the dedicated server process on node2 and accesses the sales2 database instance.

Figure 7-1 Load Balancing Actions for Oracle RAC Connections That Use SCAN

Description of Figure 7-1 follows
Description of "Figure 7-1 Load Balancing Actions for Oracle RAC Connections That Use SCAN"

See Also:

7.9.4 Understanding Shared Server Use in an Oracle RAC Database

By default, DBCA configures your Oracle RAC database with dedicated servers, not shared servers. However, if you select the shared server option when using DBCA, then DBCA configures shared servers. Oracle RAC uses both dedicated and shared server processing when shared servers are configured.

See Also: